Diastolic Dysfunction in Human Cardiac Allografts is Related with Reduced SERCA2a Gene Expression



Previous studies demonstrated that impaired left ventricular (LV) relaxation in cardiac allografts limits exercise tolerance post-transplant despite preserved systolic ejection fraction (EF). This study tested in human cardiac allografts whether the isovolumic relaxation time (IVRT), which provides the basis for most of diastolic LV filling, relates with gene expression of regulatory proteins of calcium homeostasis or cardiac matrix proteins. Gene expression was studied in 31 heart transplant recipients (25 male, 6 female) 13–83 months post-transplant with LVEF >50%, LV end-diastolic pressure <20 mmHg, normal LV mass index and without allograft rejection or significant cardiac pathology. IVRT related with the other diastolic parameters e-wave velocity (r=−0.46; p = 0.01), e/a-wave ratio (r=−0.5; p < 0.01) but not with heart frequency (r=−0.16; p = 0.4). No relation of IVRT was observed for immunosuppression, mean rejection grade or other medication. IVRT was not related with gene expression of desmin, collagen I, phospholamban, the Na+-Ca2+ exchanger, the ryanodine receptor or interstitial fibrosis but correlated inversely with SERCA2a (r=−0.48; p = 0.02). Prolonged IVRT is associated with decreased SERCA2a expression in cardiac allografts without significant other pathology. Similar observations in non-transplanted patients with diastolic failure suggest that decreased SERCA2a expression is an important common pathomechanism.


Exercise tolerance post-transplant does not return to normal in many heart transplant recipients because of cardiac, pulmonary and peripheral limitations (1,2). Cardiac limitation in these patients is related with a combination of diastolic dysfunction and chronotropic incompetence (3–6) in the presence of normal systolic left ventricular ejection fraction (LVEF) (7). Diastolic dysfunction in cardiac allografts results from impaired left ventricular (LV) relaxation, and is associated with dyspnea and limited cardiac performance when patients exercise (4) or afterload is augmented pharmacologically (3,6). Chronotropic incompetence ensues from denervation of the cardiac allograft which blunts the increase of heart rate especially in the early phase of exercise (8). Previous studies demonstrated that diastolic dysfunction in cardiac allografts is not associated with the number of rejection episodes, immunosuppressive medication or the interval post-transplant (5) suggesting that pathophysiological mechanisms not related with organ transplantation may be relevant. Mechanisms that cause diastolic dysfunction in mammalian myocardial tissue can be intrinsic to the cardiac muscle cells themselves with changes in the calcium homeostasis, or result from changes in the cardiomyocyte cytoskeleton or the extracellular matrix (9,10). Calcium homeostasis plays a most important role in the first phase of diastole, isovolumic relaxation, which is the time interval from the closure of the aortic valve to the opening of the mitral valve. While isovolumic relaxation itself does not contribute to ventricular filling, it provides the basis for the LV filling in the second phase of diastole which accounts for 80% of the cardiac output. The time interval of isovolumic relaxation (IVRT) in cardiac allografts has been reported to be prolonged (11); however, in other reports and in our own experience IVRT may vary within a large range in the single heart transplant recipient (4,12).

We hypothesized that differences in the gene expression of calcium-handling proteins should explain for diastolic dysfunction in cardiac allografts as suggested in studies in heart failure patients (13) or in senescent myocardium (14). Therefore, this study assessed gene expression of regulatory proteins of calcium homeostasis (ryanodine receptor type 2, sarcoplasmic Ca2+ATPase type 2a, phospholamban (PLB), cardiac Na+-Ca2+ exchanger, cardiac calsequestrin). Because of the relevance of cytoskletal structures and the extracellular matrix for diastolic function (9), we measured gene expression of the cytoskeletal intermediate filament protein desmin and the extracellular matrix protein collagen type I in addition. Expression of all candidate genes was assessed simultaneously in endomyocardial biopsies obtained from cardiac allografts >1 year post-transplant. This limitation should exclude the likelihood of multiple gene expression changes that accompany the dynamic cardiac allograft modeling process in the first post-operative year (15). Results from quantitative gene expression measurements were related with IVRT at the time of biopsy.

Materials and Methods

Inclusion and exclusion criteria

Included were 31 HTx recipients >1 year post-transplant. Excluded were HTx recipients with (1). histologic evidence of allograft rejection graded ≥3 according to Texas classification (16) because acute allograft rejection affects diastolic function (11); (2) LV echocardiographic EF <55%, and ≥ moderate aortic or mitral valvulopathy; (3) LV end-diastolic pressure (LVEDP) ≥20 mmHg; (4) history of uncontrolled systolic arterial hypertension with >150 mmHg on chart review; (5) heart rate <55 and >120 bpm or pace maker treatment and (6) more severe transplant vasculopathy or coronary artery lesions requiring intervention.

Patients were included after informed consent was obtained. The study was approved by the local institutional review committee (protocol number 708). Endomyocardial layers from septal or LV myocardial specimens were obtained from explanted non-failing human hearts with EF ≥55% that were not eligible for organ transplantation because of donor age (n = 3), viral infection or blood group mismatch.

Assessment of cardiac function

Transthoracic echocardiography:  Examination was performed with each participant in left lateral recumbent position using 2–5 MHz transducers and standard equipment [Acuson® Aspen, Sequoia, XP128 (Acuson, Malvern, PA, USA)]. IVRT was measured as the time from aortic closure to opening of the mitral valve with the continuous-wave Doppler positioned in the four-chamber view between the anterior mitral valve leaflet and the aortic valve. Echocardiograms were recorded for off-line analysis, measurements were the mean of three consecutive heart cycles.

Cardiac catheterization:  Endomyocardial biopsies were taken from the right ventricular aspect of the interventricular septum under fluoroscopic control. LVEDP, aortic end-systolic and end-diastolic pressures were measured with fluid-filled catheters connected to Statham transducers. Hemodynamic measurements were determined as mean of three consecutive sinus beats. End-diastolic volume and EF were calculated by the centerline method.

Endomyocardial biopsies, mRNA isolation and real-time PCR

Endomyocardial biopsies:  Specimens for gene expression analysis were obtained together with the routine biopsies when the patients were hospitalized for the annual post-transplant routine control including left and right heart catheterization and transthoracic echocardiography. Endomyocardial specimens were frozen in liquid nitrogen immediately after removal. mRNA isolation from microbiopsies was similar to Hullin et al. (17).

Gene expression in interventricular septum and free LV wall:  Layers of 2 mm thickness were prepared from the trabeculated endomyocardial regions of the interventricular septum and the free LV wall of non-failing human hearts. Biomagnetic separation was used for mRNA isolation (17).

Quantitative gene expression:  Coding sequences were cloned by RT-PCR with specific 5′- and 3′-oligonucleotide primers and reverse transcribed mRNA isolated from human non-failing LV (14). Amplified fragments were subcloned into pCR2.1-TOPO (Invitrogen®, Carlsbad, CA, USA), subcloned fragments were sequenced on both strands (MWG Biotech®, Fbersberg, Germany) and used for real-time PCR when identical to their respective GenBank sequences. Quantification in real-time PCR used standard curves amplified with 103–107 plasmids containing the subcloned coding sequences. Always, 2 biopsies/patient were measured as duplicates. For real-time PCR the iQ™ Supermix was used in the presence of a Taqman-probe; in the absence of a Taqman-probe the iQ™ SYBR Green Supermix (Bio-Rad® Laboratories, Hercules, CA, USA) was used, always followed by melt curve analysis (70°C–94°C at 0.5°C steps). Reactions were carried out on the iCycler iQ® Instrument (Bio-Rad® Laboratories AG). Concentrations of primer/TaqMan® probes were 2 μM always. Gel electrophoresis of real-time PCR reactions always visualized one single amplification product band.

Cardiac calsequestrin:  5′-primer: 5′- CTGAGCATCCTGTGGATCGAC; 3′-primer: 5′- TGTGGCCTGAATAGGTCAATCTT; positions on GenBank D55655: 1010-1030, 1098-1076. Amplification protocol: 94°C, 3 min; 57.8°C and 94°C, 30 and 20 s, 40 times; hold 25°C. Correlations: 0.997–0.999, efficiencies: 90.1–97.5%.

Phospholamban:  5′-primer: 5′- TTCTCTCGACCACTTAAAACTTCA, 3′-primer: 5′- CTGAGCGAGTGAGGTATTGGA, TaqMan® probe: 5′-6FAM- CTTCCTGTCCTGCTGGTATCATGG-Dabcyl; positions in GenBank M63603: 136–159, 202–192, 162–185. Protocol: 94°C, 3 min; 57°C and 94°C, each 30 sec, 40 times; hold 25°C. Correlations: 0.997–0.999, efficiencies: 92.3–102.1%.

Sarcoplasmic Ca-ATPase (SERCA2a):  5′-primer: 5′- TGAACCCTCCCACAAGTCTAAA, 3′-primer: 5′- CCAATCTCGGCTTTCTTCAGAG, TaqMan® probe: 5′-6FAM- CATCGTTCACGCCATCGCCAGTCA-Dabcyl; positions in GenBank M23115: 2037-2058, 2150-2129, 2122-2099. Protocol: see PLB. Correlations: 0.991–0.995, efficiencies: 92.5–106.8%.

Ryanodine receptor type 2 (RyR):  5′-primer: 5′- AAAGCCGAGGGAGAAGATGGA, 3′-primer: 5′- TCTGTTGATATGCTATGATTTTCTTCCA; positions on GenBank X98330: 13442-113462, 13587-13562. Protocol: 94°C, 3 min; 57°C 30 s, 72°C 20 s, 94°C 30 s, 40 times; hold 25°C. Correlations: 0.998–1.0, efficiencies: 90.9–91.8%.

Cardiac sodium-calcium exchanger (NCx):  5′-primer: 5′- AGACCTTCTTCCTTGAGATTGGA, 3′-primer: 5′- GCTCATTCAATAACAGGGCTTTC, TaqMan® probe: 5′-6FAM- TCTCACTCATCTCCACCAGGCG-Dabcyl; positions in GenBank M91368: 1939–1961, 2013–1991, 1989–1968. Protocol: see PLB. Correlations: 0.990–0.996, efficiencies: 94.2–104.6%.

Desmin:  5′-primer: 5′- TGACCGCTTCGCCAACTA; 3′-primer: 5′- GAGTAGCTGCATCCACGTCC; positions on GenBank U59167: 428–445, 735–716. Protocol: 94°C, 3 min; 56°C and 94°C, 30 and 20 s, 40 times; melt curve, hold 25°C. Correlations: 0.994–0.996, efficiencies: 94–99.1%.

Collagen I:  5′-primer: 5′- AAGAGGAAGGCCAAGTCGAG; 3′-primer: 5′- GATCACGTCATCGCACAACA; TaqMan® probe: 5′-6FAM- TCACCTGCGTACAGAACGGCCT-Dabcyl; positions on GenBank NM_000088: 187-206, 341-–322, 232-253. Protocol: 94°C, 3 min; 56°C, 72°C, 94°C, 20, 30, 20 s respectively, 40 times; hold 25°C. Correlations: 0.991–0.999, efficiencies: 90–98.1%.

Normalization:  Because of the unchanged gene expression of cardiac calsequestrin in the non-failing and failing human heart (17,14), its gene expression was assessed in each endomyocardial biopsy. Normalization was performed dividing the cardiac calsequestrin gene expression in the particular biopsy by the mean of cardiac calsequestrin gene expression in all biopsies.

Collagen Content

Collagen content in biopsies was estimated by color segmentation of digital images. The imaging software (Image-Pro Plus 5.1, MediaCybernetics™, Silver Spring, MD, USA) recognizes particular colors and distinct shades and digitally highlights pixels of a particular color or shade specified within the field. Fibrosis becomes blue with trichrome stain and is highlighted based on the operator's threshold settings allowing for calculation of the amount of area occupied by fibrosis within the field.


Verification of normal distribution of data was accomplished with visual inspection of histograms. Power transformations were used to normalize skewed distributions. Student's t-test for comparison of gene expression in endomyocardial layers. t-test, chi square test, Spearman rank correlation were used where appropriate. All tests were computed using the Stat View 4.57 software (Abacus, Inc., Berkeley, CA, USA).


Clinical characteristics and relation to IVRT

Patients:  Thirty-one study patients were included with a post-operative interval >1 year. No relation of IVRT was detected with recipient's age (r= 0.17; p = 0.37) or time post-transplant (r= 0.05; p = 0.79), whereas there was a trend to a positive relation between donor age at time of study and IVRT (r= 0.37; p = 0.08). No relation was observed for the mean allograft rejection score (r=−0.25; p = 0.18) as well as the number of episodes of moderate allograft rejection (r=−0.03; p = 0.86) (Table 1).

Table 1. Clinical parameters
Demography of PatientsMean±SDRangeCorrelationp Values
  1. Values are given as mean ± S.D.

  2. fT3 = free trijodothyronine, fT4 = thyroxin, TSH = thyroid stimulating hormone; ACEI = angiotensin converting enzyme inhibitor; AT-II-RA = angiotensin II type 1 receptor antagonist.

  3. p values <0.05 were considered statistically significant.

 Gender (% males)77 
 Age at catheterization (year)541227–720.170.37
 Time post-transplant (days)14741035374—53860.050.79
Donor Hearts
 Age (at cathetherization, year)411519—660.370.08
Clinical Variables
 Episodes of moderate rejection (mean)1.31.70–4−0.250.18
 Texas score (mean)–2.7−0.030.86
 fT3 (pmol/L)–
 fT4 (pmol/L)–
 TSH (pmol/L)–5.0−0.160.39
 Cyclosporine (n = 29)
  Mean daily dosage (mg)18390125–4000.240.20
  Dosage day of cath (mg)1516693–2970.140.47
  mean blood level (ng/mL)18224136–2220.170.37
Tacrolimus + Sirolimus (n = 2)
Prednisone (n = 25)
 Mean daily dosage (mg)173.49.8–25.3−0.050.80
Azathioprine (n = 7)
 Mean daily dosage (mg)66.817.227—125−0.020.93
Mycophenolate (n = 18)
 Mean daily dosage (mg)25125421740–3250−0.260.16
 Mean blood level (mg/L)–3.1−0.210.25
Cardiovascular Medication
 β-blockers (n = 2) (%)6.5 
 Calcium antagonists (n = 24) (%)81 −0.110.56
 ACEI and AT-II-RA (n = 22) (%)71 −0.300.10

Medication:  Immunosuppressive therapy was adapted to the individual course of allograft rejection in every patient. IVRT was not related with mean daily dosages of cyclosporine (r= 0.24; p = 0.20), azathioprine (r=−0.02; p = 0.93) or mycophenolate (r=−0.26; p = 0.16), or mean blood levels of cyclosporine (r= 0.17; p = 0.37) or mycophenolate (r=−0.21; p = 0.25). Two study participants received β-blockers; 24 participants were on calcium antagonists (diltiazem) (r=−0.11; p = 0.56), and there was a trend for a negative relation of IVRT with neurohumoral blockade of the renin-angiotensin system (r=−0.3; p = 0.10) (Table 1).

Hemodynamic parameters

IVRT was not related with heart frequency (r=−0.16; p = 0.4). No correlation of IVRT was identified with the echocardiographic LVEF (r= 0.01; p = 0.97) or LV mass index (r= 0.0; p = 0.99). However, IVRT related negatively with the two other diastolic parameters e-wave (r=−0.46; p = 0.01) and the ratio of e/a-wave (r=−0.5; p < 0.01), but not the e-wave deceleration time (r= 0.2; p = 0.28) which reflects the compliance of the LV in the 3rd phase of diastolic filling. No correlation was identified for the LVEF as determined by the centerline method at cardiac catheterization (r=−0.1; p = 0.59) or LVEDP (r=−0.29; p = 0.11); however, there was a trend for a negative relation with invasive pressure decay during isovolumic relaxation (r=−0.33; p = 0.07) (Table 2).

Table 2. Hemodynamic parameters
  1. Values are given as mean ± S.D. bpm = beats per minute; EF = ejection fraction, LVMI = left ventricular mass index; IVRT = isovolumic time of relaxation; eDecel-time = e-wave deceleration time; LVEDP = left ventricular enddiastolic pressure; dp/dt = isovolumic left ventricular pressure measured invasively.

  2. p values <0.05 were considered statistically significant.

Pulse (bpm)801260–93−0.160.4
Transthoracic echocardiography
 LVMI (g/m2)1072551–1520.000.99
 e-Wave (m/s)–1.3−0.460.01
 a-Wave (m/s)–
 eDecel-time (ms)16035110–2370.20.28
Cardiac catheterization
 LVEDP (mmHg)1042–17−0.290.11
 EF HK (%)68655–83−0.10.59
 dp/dt (mmHg/s)2043464945–2875−0.330.07

Gene expression

Interventricular septum and free LV wall:  In endomyocardial layers of septal and LV free wall specimens obtained from non-failing hearts gene expression was not significantly different for the genes examined: SERCA2a (n = 8; 2.13 ± 4.2 × 106 vs. 9.05 ± 1.58 × 105 copies; p = 0.29), PLB (n = 5; 1.72 ± 0.28 × 107 vs. 8.63 ± 1.29 × 106 copies; p = 0.26), sodium-calcium exchanger (NCx) (n = 5; 5.16 ± 0.85 × 106 vs. 8.21 ± 1.23 × 106 copies; p = 0.28), RyR type 2 (n = 5; 8.33 ± 1.65 × 108 vs. 4.93 ± 1.12 × 109 copies; p = 0.42), desmin (n = 5; 7.73 ± 1.48 × 103 vs. 4.07 ± 0.9 × 104 copies; p = 0.41) and collagen type 1 (n = 5; 4.45 ± 0.65 × 105 vs. 8.36 ± 1.35 × 105 copies; p = 0.31).

Gene expression in endomyocardial biopsies related with IVRT:  Gene expression of the ryanodine receptor type 2 was not related with IVRT (r= 0.17; p = 0.37), as was gene expression of the cardiac sodium calcium exchanger (r=−0.15; p = 0.41) and PLB (r= 0.14; p = 0.44). For SERCA2a gene expression a strong negative correlation with IVRT was observed (r= 0.48; p = 0.02) (Figure 1).

Figure 1.

Gene expression of regulatory proteins of cardiomyocyte calcium homeostasis. Scatter plots show gene expression of (A) sarcoplasmic Ca2+-ATPase (SERCA2a), (B) phospholamban (PLB), (C) Na+-Ca2+-exchanger (NCx) and (D) ryanodine receptor type 2. IVRT: isovolumic time of relaxation in milliseconds (ms).

Gene expression of the intermediate filament protein desmin was not related with IVRT (r= 0.01: p = 0.94), nor was gene expression of the extracellular matrix protein collagen type I (r=−0.25; p = 0.18). Color segmentation of digital images of trichrome stained endomyocardial biopsies demonstrated that interstitial fibrosis was not related with IVRT (r= 0.09; p = 0.63); however, collagen type 1 gene expression related with interstitial fibrosis (r= 0.51; p = 0.02) (Figure 2).

Figure 2.

Gene expression of collagen I and desmin. Scatter plots of normalized gene expression of (A) collagen type I and (B) desmin. (C) Collagen content in trichrome stained endomyocardial biopsies estimated by color segmentation of digital images (pixels counted). IVRT: isovolumic time of relaxation in milliseconds (ms).


Our study demonstrates that prolongation of isovolumic relaxation time in cardiac allografts is related with decreased SERCA2a gene expression; however, no direct association of decreased SERCA2a gene expression with any transplantation specific parameter was identified.

Isovolumic relaxation is most relevant for cardiac output because it provides the basis for the rapid LV filling in the second phase of diastole which accounts for almost 80% of the LV filling. Isovolumic relaxation results from the active reduction of the cytosolic calcium concentration by (1) the predominant reuptake of calcium ions into the sarcoplasmic reticulum and (2) the extrusion into the interstitial space by the sarcolemmal NCx disturbances in intracellular calcium homeostasis have direct effect on this first phase of diastole.

Measurement of isovolumic relaxtion

The gold standard for assessment of isovolumic relaxation is the microtip-catheter with digital measurement of the LV pressure decay. Because measurements of IVRT by microtip-catheters and echocardiographic measurement of isovolumic relaxation relate well when LV end-diastolic pressure (preload) is below 20 mmHg (18), inclusion of study participants was limited accordingly. This concept was confirmed by the good correlation of IVRT with the e-wave velocity which is the peak LV filling rate in the second diastolic phase and depends directly on IVRT only when preload is comparable. In addition, IVRT related well with the e/a-wave ratio which is the ratio of LV filling rates in the second (LV peak filling) and fourth diastolic phase (left atrial contraction) that decreases with increasing impairment of diastolic function. Furthermore, there was a trend for a positive correlation of the echocardiographic IVRT with the decay of pressure in isovolumic relaxation measured by fluid-filled pigtail catheters (r=−0.33; p = 0.07), although fluid filling dampens LV pressure decay measurements. Altogether, these findings maintain the adequate assessment of isovolumic relaxation by echocardiographic measurement of the time interval from the closure of the aortic valve to opening of the mitral valve in our study population.

Isovolumic relaxation and other clinical parameters

IVRT is related directly to heart rate with parallel acceleration of IVRT with heart rate (19). The importance of this physiological interdependency of IVRT and heart rate is reflected by the observation of a constant ratio of IVRT to heart frequency in infancy and adolescence (20). This explains why impaired myocardial relaxation with prolonged IVRT may not result in production of elevated LV filling and clinical signs of diastolic dysfunction when the heart frequency is low, but will result in elevated LV filling pressure when the heart frequency is high (19). IVRT in our study participants varied within a broad range similar to observations in the literature (4,11); however, no relation with heart frequency was observed indicating that the impact of heart frequency on IVRT was not relevant. Similarly, IVRT was not related with immunosuppressive medication, mean histological grade of allograft rejection or the number of moderate allograft rejections and the time post-transplant — similar to previous reports (4,11). A substantial number of study patients received the calcium antagonist diltiazem which has both lusitropic and negative chronotropic effects. Whereas diltiazem does not alter parameters of diastolic function in patients with minor coronary artery disease and normal LVEF (21), it may improve diastolic function in patients with hypertrophic cardiomyopathy (22). In our study population, diltiazem medication was not related with IVRT or heart frequency suggesting that diltiazem intake did not impact significantly on IVRT. Furthermore, LV mass index in all study participants was normal which excludes a relevant influence of diltiazem on relaxation as demonstrated for LV hypertrophy. In our study patients, IVRT was related with donor heart age but this relation did not reach significance. Nevertheless, this suggests that impairment of ventricular relaxation in transplanted hearts may be related with age similar to animal models and humans (23,14). Altogether, no significant relation was identified between IVRT and medication or other clinical parameters which provided the basis for correlating gene expression with the IVRT.

Gene expression in endomyocardial biopsies

Myocardial specimens were obtained from the right ventricular aspect of the interventricular septum which is an integral part of the LV. When we tested whether gene expression in endomyocardial layers of non-failing human myocardium corresponds between the right ventricular part of the interventricular septum and the free LV wall, no significant differences were observed with respect to the expression of the candidate genes. As our study was restricted to the analysis of gene expression in endomyocardial biopsies with a size of one to few milligrams, we used real-time polymerase chain reaction for quantitative assessment. While this allowed for analysis of the different candidate genes in the same biopsy at same time, it restricted the investigation to mRNA measurements. However, in endomyocardial layers gene expression at the mRNA levels correspond to their respective protein expression in non-failing and failing human heart, at least for SERCA2a, NCx and cardiac calsequestrin (24). Other studies did not observe this consistent relation of SERCA2a mRNA and protein expression in homogenates of human LV non-failing and failing heart specimens (25). However, SERCA2a gene expression is different in endo-, mid- and epicardial layers with corresponding changes at the mRNA and the protein level (24) suggesting that myocardial specimens with variable portions of myocardial layers may result in the described incongruency of SERCA2a mRNA and protein expression.

Our gene expression studies in endomyocardial biopsies demonstrated an inverse relation of SERCA2a gene expression with IVRT, whereas PLB and NCx gene expression remained unchanged. This suggests that reuptake of cytosolic calcium ions into the sarcoplasmic reticulum may be reduced due to the decreased number of SERCA2a molecules. Conclusions based on the relation of gene expression with cardiac function should be verified in functional investigations always which were not feasible within our study where two additional biopsies were obtained for research purposes already. In the literature, several studies provide evidence for a relation of SERCA2a mRNA and protein gene expression with impaired relaxation: mice heterozygous for a null mutation in the SERCA2a gene show concordant reductions of SERCA2a at the mRNA and the protein level and prolongation LV relaxation due to slowed maximal velocity of Ca2+ reuptake into the sarcoplasmic reticulum (26). Likewise, reduced sarcoplasmic Ca2+ reuptake has been observed in cardiac allografts suggesting a reduction in SERCA2a activity but in this investigation diastolic function in the respective cardiac allografts was not studied (27). Atrial trabeculae obtained from senescent human myocardium demonstrate a reduced reuptake of calcium into the sarcoplasmic reticulum, and impairment of diastolic relaxation in that study was associated with reduced SERCA2a gene expression (14). Furthermore, diastolic function remains preserved in patients with systolic end-stage cardiomyopathy when SERCA2a gene expression is unchanged, whereas reduced SERCA2a gene expression is associated with diastolic dysfunction (13). SERCA2a activity is under the inhibiting regulatory control of PLB. Gene expression of PLB remained unchanged in our study patients suggesting a decrease in the ratio of SERCA2a to PLB which may increase the inhibitory action of PLB similar to studies in PLB overexpression in mice (28). In the non-failing heart β-adrenergic mediated phosphorylation relieves the inhibitory action of PLB on SERCA2a activity; however, intracellular β-adrenergic signaling is blunted in the cardiac allograft (29) suggesting that the ratio of PLB to SERCA2a may play a more important role post-transplant.

Clinical implications of SERCA2a gene expression and diastolic dysfunction

In the human heart, SERCA2a gene expression is under the regulatory control of the thyroid hormone (30). In our study population, thyroid hormone levels were not related with SERCA2a gene expression but this association may offer a therapeutic strategy in the case of hypothyroidism in heart transplant recipients with prolonged IVRT.

Cyclosporine is a known direct isoform-specific blocker of skeletal muscle SERCA type 1 (31) without pharmacological effect on cardiac SERCA2a activity. This may explain why no association of cyclosporine was observed with IVRT in our study population. On the other hand, stretch-induced up-regulation of SERCA2a gene expression is blocked indirectly by inhibition of the calcineurin/NFAT pathway in the human heart (32) indicating that cyclosporine plays an indirect role and discontinuation of cyclosporine medication may offer a therapeutic option therefore. Because of the underlying gene expression changes associated with prolonged IVRT adequate chronotropic control is a first effective step to control symptoms of diastolic failure and future studies have to demonstrate whether adequate frequency control may improve exercise tolerance and increase peak VO2 in patients with impaired diastolic function.

Limitations of the study

This study should be interpreted with several important caveats. First, in the cardiac allograft gene expression in the right ventricular side of the interventricular septum may not be representative for expression of the respective candidate gene in the free LV wall despite unchanged expression of respective candidate genes in the non-failing heart. Second, we chose to study gene expression in endomyocardial biopsies from the interventricular septum because of access and safety of tissue removal which on the other hand limits the amount of tissue obtained. This restricted gene expression analysis to measurements at the mRNA level in order to study the seven different candidate genes. Therefore, in our study we relate gene expression at the mRNA level directly to functional measurements which can be debated because mRNA levels may not correspond to their respective protein levels and, in addition, protein levels themselves may not relate directly to function because of the intracellular regulatory mechanisms such as posphorylation and dephosphorylation. However, the pathophysiological relevance of our principal finding is suggested because of the analogous findings in other models of diastolic dysfunction.


We thank the Swiss Heart Foundation, the Inselspital Foundation and the Katharina Huber-Steiner Foundation for their financial support. Our special thank is directed to the cardiothoracic surgeons of the Cardiovascular Center Bern.